Insights into the Evolution of Chemical Degradation in Fuel Cell Membranes Using 4D in Situ Visualization

Abstract

Ionomer membrane degradation during fuel cell operation is conjointly produced by three different modes, namely, thermal, mechanical, and chemical. Amongst these, chemical degradation is generally considered a major contributor to lifetime limitations.[1] Chemical degradation of the ionomer membrane is caused by the formation of radical species, such as hydroxyl (OH•) and hydroperoxyl (HOO•) groups, which causes polymer chain scission and unzipping leading to membrane thinning and eventual rupture and shorting events. Furthermore, the chemical degradation deteriorates the mechanical properties of the membrane which dramatically accelerates its overall degradation rate. Scanning electron microscopy (SEM) based studies have been conventionally used to conduct two-dimensional characterization of degradation-induced structural features in fuel cell membranes; however, SEM imaging is inherently destructive and thus prohibits any tracking of structural changes over time. Laboratory-based X-ray computed tomography (XCT) is as an alternative imaging technique, which has not only enabled three-dimensional (3D) membrane failure analysis revealing novel insights on membrane failure [2,3], but has also facilitated studies on damage growth characterization, [4] and more recently on evolutionary aspects of mechanical membrane degradation through identical-location imaging. [5] In the present work, an XCT-based 4D in situ imaging method, featuring three dimensions in space and one dimension in degradation time, is applied to investigate the evolution of pure chemical membrane degradation through the use of a custom designed fixture housing a fully operational small-scale fuel cell. [6] Membrane electrode assemblies (MEAs) with three distinct membrane types, namely, i) non-reinforced, ii) mechanically reinforced, and iii) chemically and mechanically mitigated membranes, are separately examined and compared. The MEAs were subjected to open circuit voltage (OCV) hold under high temperature (75°C) and low humidity (30% RH) with 0.3 slpm H2 and 0.5 slpm air supplied to anode and cathode, respectively. Various in situ electrochemical diagnostics were periodically performed to monitor MEA and membrane health. XCT-based 3D datasets were obtained at various stages of degradation to facilitate identical location tracking and analysis of membrane morphology as a function of degradation time and enable the characterization of damage evolution in the three membrane types. Membrane thinning and other damage features within all three membranes are comprehensively examined from various perspectives by studying planar and cross-sectional views extracted from the 3D XCT datasets. Preliminary observations show that electrode-shorting under land regions was the key failure mode for membranes without chemical mitigation, i.e., non-reinforced and mechanically reinforced membranes (Fig. 1). In contrast, the chemically mitigated membrane did not fail up to 850 h of operation, highlighting the scavenging effect against the radicals. No membrane pinhole or crack development was observed in any of the three membranes, which is consistent with previous 3D post mortem results under pure chemical degradation.[3] The present 4D in situ imaging approach further revealed locally amplified membrane thinning preceding the eventual formation of shorts due to absence of membrane material and associated loss of electrode separation. The observed phenomenon further suggests that mechanical stresses, even in the absence of wet/dry cycling, may influence the local shorting events. Overall, the new findings from this work demonstrate the distinct advantage of XCT technology towards improving the fundamental understanding of membrane degradation by capturing critical failure modes and mechanisms at their different developmental stages. Acknowledgements This research was supported by the Natural Sciences and Engineering Research Council of Canada, Canada Foundation for Innovation, British Columbia Knowledge Development Fund, and Ballard Power Systems through an Automotive Partnership Canada grant. This research was undertaken, in part, thanks to funding from the Canada Research Chairs program. W. L. Gore & Associates, Inc. is acknowledged for additional support.